Independent LLM benchmarks

💰

LLM Economics

Compare hardware, models and inference setups by operating cost, token output and total cost of ownership.

💸 Operating cost🖥️ TCO⚡ Tokens / euro🔋 Tokens / kWh

📊 Overview 🎮 GPUs 🧠 Models ⚖️ Compare 📚 Methodology

Settings are stored in your browser and shared via the URL. Power draw is estimated from GPU TDP × count since no per-run measurement exists. Methodology →

GPU economics

9 GPU configurations

Aggregated per configuration (incl. GPU count) – economics computed per run, then AVG/MIN/MAX. Number after the GPU = runs · models.

GPUby GPU

NVIDIA RTX PRO 6000 Blackwell Workstation Edition 1.443,9 tok/sNVIDIA GeForce RTX 5090 1.386,8 tok/sNVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition 1.238,3 tok/sNVIDIA GeForce RTX 3090 Ti 901,9 tok/sAMD Radeon AI PRO R9700 340,5 tok/sNVIDIA GB10 (DGX Spark) 174,2 tok/sAMD Radeon 8060S Graphics 173,3 tok/s

GPU configurations chart

NVIDIA GeForce RTX 3090 Ti

EUR 0.042

AMD Radeon 8060S Graphics

EUR 0.058

NVIDIA GB10 (DGX Spark)

EUR 0.067

NVIDIA RTX PRO 6000 Blackwell Workstation Edition

EUR 0.035

NVIDIA GeForce RTX 5090

EUR 0.035

NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition

EUR 0.87

3× AMD Radeon AI PRO R9700

EUR 0.22

3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition

EUR 0.061

AMD Radeon AI PRO R9700

EUR 4.49

Bars per configuration – switch between cost per 1M tokens, tokens per euro and throughput. Cheaper/faster = longer bar.

Over 2 years projected

Projection per GPU configuration for the selected period: electricity cost and generated output tokens (per-run AVG × effective operating hours). Adjust period/utilization above.

Cost vs. generated tokens (2 years)

Cost over 2 years

Generated tokens over 2 years

Cost & tokens grow linearly with runtime. Power = GPU TDP × count (estimated).

#	GPU configuration	Gen tok/s i	Cost/h	EUR / 1M tok i	Tok/€	Tok/kWh	TCO	Data
1	NVIDIA GeForce RTX 3090 Tii24 GB VRAM · EUR 999 · 450 W TDP	389.9 8–902	EUR 0.14	EUR 0.042	24.05M	7.21M	EUR 2,365	estimated (TDP)
2	AMD Radeon 8060S Graphicsino price · 120 W TDP	67.3 7–173	EUR 0.036	EUR 0.058	17.33M	5.20M	EUR 631	estimated (TDP)
3	NVIDIA GB10 (DGX Spark)i128 GB VRAM · no price · 140 W TDP	86.6 32–174	EUR 0.042	EUR 0.067	14.93M	4.48M	EUR 736	estimated (TDP)
4	NVIDIA RTX PRO 6000 Blackwell Workstation Editioni96 GB VRAM · EUR 8,999 · 600 W TDP	503.1 9–1,444	EUR 0.18	EUR 0.035	28.88M	8.66M	EUR 3,154	estimated (TDP)
5	NVIDIA GeForce RTX 5090i32 GB VRAM · EUR 2,299 · 575 W TDP	635.7 3–1,387	EUR 0.17	EUR 0.035	28.94M	8.68M	EUR 3,022	estimated (TDP)
6	NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Editioni96 GB VRAM · EUR 8,499 · 300 W TDP	28.8 29–29	EUR 0.090	EUR 0.87	1.15M	345.60K	EUR 1,577	estimated (TDP)
7	3× AMD Radeon AI PRO R9700i96 GB VRAM · EUR 3,897 · 900 W TDP	73.8 2–341	EUR 0.27	EUR 0.22	4.54M	1.36M	EUR 4,730	estimated (TDP)
8	3× NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Editioni288 GB VRAM · EUR 25,497 · 900 W TDP	45.1 5–1,238	EUR 0.27	EUR 0.061	16.51M	4.95M	EUR 4,730	estimated (TDP)
9	AMD Radeon AI PRO R9700i32 GB VRAM · EUR 1,299 · 300 W TDP	5.6 6–6	EUR 0.090	EUR 4.49	222.80K	66.84K	EUR 1,577	estimated (TDP)

MIN = lowest cost (highest speed / lowest draw), MAX = highest cost. Every cost metric is bound to a real run (computed per run, then aggregated). Prices = GPU acquisition (unit price × count).